Pyra-news January 2016

pyra-news@pyra-handheld.com

1 participants
4 discussions

Let's talk Colors
by Michael Mrozek 21 Jan '16

21 Jan '16

I hope you're not fed up with newsposts yet... but that will be the last one during my Greek-visit :) Okay, before we start, one more video of the current CNC Milling: https://youtu.be/nsDE2f4E8u0 First you can see the final step of the tooling for the keyboard part. The copper slowly cuts into the metal using electricity. That takes quite a while. The second thing you can see is basically the LCD frame being "stamped" into the metal... you can imagine how long that takes. It's a slow process, but the only way to do it accurately without any issues. The last thing you can see is a typical CNC drill that mills into the metal block. That ALSO takes a while, as you can it is moving slowly. Yes, creating molds suitable for plastic injection is a quite a lengthy job, if done right. Altogether, the Pyra will have 7 molds. Okay, let's talk about the colors now! -------------------------------------- So, the material we're using for the Pyra is the transparent one you've seen in the video yesterday... nearly unbreakable Polycarbonate. How does the color get into the case? Well, while the process itself is easy (you mix some powder together with the plastic pellets), this also needs quite a bit of preparing. Depending on the plastic you use, you also need to use different color types. Different types of plastic have different melting temperatures, so the powder you use has to be the correct one as well. Otherwise, the color wouldn't melt or would burn. Both would lead to horrible results. Also, you need to find the perfect mixture to make the color look good. This also depends on the wall thickness of the plastic part itself... And as there are so many different types of colors available, the initial selection is very limited. Usually, you get the typical industrial colors (bright red, yellow, green, etc.). For anything decent looking (a nice dark blue or red), you often need to get a customized color, made especially for you. Additional stuff like a metallic look or the scratch resistant coating will also be included within that custom color. Of course, custom colors need to be bought in higher quantities as well, and they take a while (3 - 7 weeks) before you get them from the manufacturer. We will make samples in 5 different colors. To save costs and speed things up, we've decided to use 3 readily existing ones that look good and make 2 custom made ones. None of them will have any special coating or anything added, as that adds more costs and will take longer. The mass produced case will have a nice, decent color with matte scratch resistant coating and maybe a tiny little bit of metallic (just to make the color look good, I'm not aiming for a metallic look!) It could very well be we'll produce more than one color for mass production, but they will all have the matte scratch resistant coating. As the plastic itself is colored, it can never rub off (though I doubt it would even rub off with that coating :D) Okay, so here is how we proceed from now on: -------------------------------------------- The mold for the lower part of the basement (with the battery compartment) is already finished. The battery lid, LCD frame and keyboard main part mold will be finished end of next week. They will then produce transparent physical testing versions of those parts and send them to me so I can test that everything fits and make sure the shoulder buttons, etc. are working fine. This will only be a physical sample, it will use a cheap plastic and fast production run, it will not look pretty (so do not be scared when you see the pictures in 2 weeks :)) Once I confirm the samples to be working, these parts will immediately be produced with the chosen colors. These will look better than the physical sample, but not as good as the mass produced one as the scratch-resistant coating is still missing. Of course, while I'm testing this part, the last two molds, the logo and the LCD backside will be finished (approx. 1 week as well), there won't be a break in the mass production. They will also be produced in the chosen sample colors and then it's time to assemble the units for the prototype customers (who have given me their three favorite colors). As soon as I make the decision about the final color (maybe you can help here at the boards ;)) or final colors, the custom color mix will be ordered, including the scratch-resistant coating any maybe some slight metallic look. Once the colors have arrived, mass production can take place in any quantity. 2000 case sets would take approximately 1 week. While all this is happening, Nikolaus and I will continue to work on the PCB testing and assembling as well as setting up the mass production with GC and ordering the parts once the hardware is confirmed to be working. Want to leave feedback and look at the pictures? Visit: https://pyra-handheld.com/boards/threads/lets-talk-colors.76884/

1 0

You'll hate this post. And you'll love this post.
by Michael Mrozek 21 Jan '16

21 Jan '16

First day here in Greece is over. Quite a shocking start, but a lot more confidence in the afternoon. What happened? Read on :) With spicy pictures (only on the boards) and news :) Let's first start with the shock. You'll hate this. I hated this: I can't show you a finished case set. I didn't get one. All I got was one ugly looking basement part. Looks similar to the first Pandora case without paint we got... what a start of the day. And it was the start of me learning how plastic mould injection is working in detail. It's quite complex. But I've learnt so much new things today! And I'll try to tell it to you as good as I can :) So, why did they tell me they've got a full set for me? Because they have. Not a full set of finished cases - but a full set of the molds. I simply understood that wrong. Probably combined with the anticipation of getting a final case :D What they have is the first step of the molds, to be exact: They're done, you could produce cases with them them, but they have no tweaks applied yet and no surface modification. This probably is comparable to the final molds we got for the Pandora: Physically, they were correct, but they had shrinking marks (as they weren't tweaked yet) and they didn't have a surface modification, so they looked ugly (and we decided to paint them...) So our step 1 here was basically the final step of the Pandora molds. Here is a basic overview how plastic mold injection works. If you're not interested how that works, skip the part between the lines :) --------------------- 1. 3D Design ------------ The first thing is the 3D design. CAD. That's the 3D model where we got the renderings from. 2. Creating the molds --------------------- Then you need to turn these models into molds. And THAT'S a very tricky part. If you think it's comparable to 3D printing: Nope, it isn't. Totally different. A mold is basically like a box: You close it and pump hot plastic into it. It then cools off and gets hard. Then your plastic part is done. So the mold needs to be a negative (inverted version) of the plastic part, enclosed in two metal parts. No problem for simple parts (like the LCD frame of the lid). It gets trickier with more complex parts though, like the keyboard part of the case: There are holes in the area where the hinge is. Holes inverted mean: The Metal needs to float in the air! You cannot create such a complex part with fine details (metal still needs to be CNC-milled, after all), so you need to create another part for the mold that you can put inside when you assemble the mold. With complex parts, the mold is a bit like a puzzle, with various inlays to create the inside-holes and special shapes. That's also needed for the clips of the battery lid, for example. When 3D printing, you have your 3D data in the computer and tell it to print it out. With mold creation, you need to first think how to achieve what you need and then create the molds including various additional tools with a CNC mill. However, that's not the only complex thing. CNC milling the steel molds is not accurate enough for fine details like the grid for the keyboard part. You need to create ANOTHER tool to do slow, and slowly cut them directly into the hot steel (with permanent oil cooling). You can see that in the picture here (and also in the video below): Creating ONE row of these key rows takes approximately five hours. In the end, however, you will have nice, clean and perfect steel mold that can be used for about 1 Million cases... Now I can totally understand why mold production is very cheap in China, but very expensive in most parts of Europe: It's a LOT of labour work - and wages are usually very high in Europe. A full case set like the one from the Pyra takes a couple of months until it's finished. 3. Testing and tweaking the molds --------------------------------- Creating the molds is already a complex thing, but it doesn't automatically mean they will be perfect right away. Similar to tweaking the machine file and PCB layout for pick & place mass production (which is what we did the pre-production run at Global Components for), you also need to test and tweak the molds. Producing plastic works like this: Melted plastic is pumped into the mold and there it cools off. The important thing is that it EVENLY cools. Otherwise, various problems can occur: Color differences, areas where it automatically can develop cracks after a while (remember the Pandora?) shrinking marks or slightly bent parts. Imagine, for example, that you have a huge peg in the case. If that one cools off faster than the ground plane, then the weight pushes it down and affects the still melted plastic - that's something that can be seen on many cheap plastic parts. Also, when cooling off, the plastic can also shrink a bit, so the size is a tiny bit different than in your model. This is also something you need to test and then fix in the mold. I am almost sure that was the problem with the LCD frame of the Pandora, the reason why the lower side has such a gap: The outer link shrank a bit and hasn't been tweaked, therefore the frame is being pushed together from both sides resulting in the gap. While the mold flow can be simulated to some extents, it's nigh impossible to accurately simulate it. So you need experience (to choose the correct places to inject the plastic before creating the mold) and need to analyse the produced parts, doing small tweaks to the molds or the machine program (you can change the pressure and heat for the plastic injection, which also improves things). 4. Creating the material surface -------------------------------- After everything has been tweaked so that the molds are basically ready to produce the plastic parts, you should improve the look of the plastic by applying some material surface to the mold. Otherwise, the plastic will simply look plain and glossy (and totally ugly), and it easily receives scratches. By applying material surface, you can create a smooth surface on the plastic that makes it look matte and it even covers smaller issues (small shrinking marks you can't prevent). We will use a very smooth surface for the Pyra, which gives it a nice, matte look. You get this effect by eroding the mold. The finer you want it, the longer it takes. That part needed about 30 hours, just for the material surface! 5. Plastic injection! --------------------- Finally, once the molds are finished, the plastic injection can happen! That's pretty straightforward (as it usually is when you do mass production, mass production has to be well prepared, simple and reliable to prevent huge failures). The plastic we use is a fully transparent one (the same one you can see in the video below) and by adding some pulverized color, you create the color you want. Once the molds are finished, thousands of case sets can easily be produced within a few days. Except you want... 6. Special treatment afterwards ------------------------------- Some things like a scratch-resistant coating can directly be applied at FormAction (so that won't need much more), but some effects are being done with a partner company, for example, textures on the case like these: Yes, we will produce some units looking like these - but those will be special editions, as it add approximately 30 EUR to production costs on top! It's cool seeing what's possible though! So much for our small introduction to mold injection... let's continue with our plans. --------------------------------- Okay, so as mentioned, the basic molds are finished, currently they are tweaking them to make sure they don't have any issues as well as applying smooth surfaces on top. One mold (the lower part of the basement, with the battery compartment) will be completely finished tomorrow (though I don't know if it's still in time so I can at least take one sample with me). The mold for the top side (the keyboard area) will be finished until end of next week. Then we'll have a complete base to test. Next will be the shoulder buttons (so we can test those, as these are the parts with the lowest tolerances), and after that, the lid will be finished. Then we got a full set! They will produce samples shortly after each mold is finished, so I will get regular shipments of new sample parts regularly until the set it complete. The samples I get will be dark transparent which helps finding out any issues, as you can see if some parts don't fit properly. When I've tested and approved everything, the cases will be produced in various colors. We've chosen about 10 colors to try out (so the prototype buyers will get one of their favorite case colors as well!) and then it's time to choose the mass production colors... then we're done here. We've also decided on the plastic we'll use. The base plastic is transparent, but by adding pulverized color, it will be completely opaque in that color. The material is very flexible and durable. If you bend it, it doesn't break but immediately move back into it's original position. According to the manufacturer, it is bulletproof (though I will NOT suggest you test that...) I've captured a bit of the durability test using my camera... though I missed the best part: someone jumping on that plastic part, and the small hook you can see on the case didn't break. Should be good enough for the battery lid, eh? BTW: A quick physical test of the ugly-produced base part was successful. I could put in the mainboard, insert a battery (and got some LEDs light up, so it works) and opening / closing the simcard as well as the MicroSD-card slot worked as well. Here is the small video, first showing a bit of the production of the keyboard mold as well as the durability test of the plastic we've chosen. https://youtu.be/q0X97nV2yaM Phewww, what a day. I am sure you can imagine how shocked I was at the beginning, learning that I won't have a full set to assemble here, and the one part I've seen was ugly... but learning about all that above was interesting. And it was really a good thing I came here, to decide about the material surface, the plastic and the color! I hope you enjoyed the newspost, even though it was looooong and too much to read, probably :) Want to leave feedback and look at the many pictures? Visit: https://pyra-handheld.com/boards/threads/youll-hate-this-post-and-youll-lov…

1 0

Greetings from Greece and Bavaria
by Michael Mrozek 19 Jan '16

19 Jan '16

Well, here I am. I just arrived at the hotel in Greece after having some great lunch with FormActions CEO and some other workers... and some coffee by the sea side :) But now it's time to post what you're waiting for... the latest status update. The status of the PCBs ---------------------- Well, I've visited Nikolaus yesterday to pick up a PCB set for the prototype case production run here in Greece. He showed me what you can also see on this video here (you probably have seen it already): https://www.youtube.com/watch?v=gk6ZB-MBezY So, what is it that you can see on this video? This is a prototype setup that has fully been booted into X. It boots off the CPU PCB (right now, that's Nikolaus testing PCB, but there's no difference than with the normal ones, only that he can debug a bit more here), the display works (incl. touchscreen), the rotator works, the LEDs work, all the hardware on the PCB physically works (means: we haven't made extensive tests i.e. with Wifi connections or usage, but the module is there and answers when we talk to it). The LCD doesn't show anything really useful, as it's Nikolaus' testing card, but it booted into X (LXDE). The only connection going into the set is the Micro USB Power cable. Nothing else. So that's the very first really working booted up prototype (except for the missing case) :) Oh, and it also works with a battery, but it's hard to make a video with that, without having a case that keeps the battery in place. So YES, we're getting there :) The low-level software is still not yet finished (only almost), some things still need to be implemented (some power stuff), so I couldn't take a full working set with me to Greece, but hopefully, when I come back, we can start building them soon :) The status of the Case ---------------------- I am here in Greece, currently in my hotel, and I'll be at FormAction (the case manufacturer) the next two days. The moulds are all finished, they already produced a sample case (and I can't wait to see it!). We're doing multiple things these two days: * Check that everything fits physically. Especially test the shoulder buttons, keymat and DPad. * Check different material samples and coatings. We're looking for a matte, scratch-resistant plastic which does not show every fingerprint in the world. * Check what colors are available and which one we can use. * Check lighting for the logo. * Check the seating and enclosure of the speakers * Check whether maybe it makes sense to use some rubber coating or similar on the nub surface So, yes, we'll be busy for the next few days. I will take one sample home with me, and different colors and materials will be shipped to me later. As usual, I'll keep you informed and will make some pictures :) Want to leave feedback? Visit: https://pyra-handheld.com/boards/threads/the-flight-has-been-booked.76833/

1 0

The flight has been booked!
by Michael Mrozek 08 Jan '16

08 Jan '16

The time has come. The flight to Greece has been booked. I'll leave on Tuesday morning, January 19th and will come back Friday noon, January 22nd. That's two full days to test out and tweak the cases! Exciting times, isn't it? Oh - you're also asking what Nikolaus has been doing during the last week? Okay, here's your answer: Fixing some small hardware bugs while working on the boards. The Display Board ----------------- None of the display boards was working when we got them - something that cries out for some hardware bug. Nikolaus debugged it for a day and luckily it was a simple fix: To save some costs, we replaced a power supply part for the display board with a different ones but that one did provide a bit lower voltage. Easy fix: Change the resistor. Now everything is working fine here, including the rotator chip using our own display cable. Oh, and as Nikolaus was working on the board anyways, he checked the sync output generated by the SSD and could confirm it sends sync signal each frame which we can use for tear free rotation with some software tweak. The mainboard and CPU board --------------------------- There was another slight issue with the mainboard power circuit which also could be fixed by simply choosing a different resistor. What happened: Booting the system only worked using a battery or through the Console / Charger port. When trying to power it through the USB OTG port, it didn't get enough power. Why's that? Well, usually, there's 1,8V set on the USB OTG port, so that a computer connected to it recognizes it as a high-power device. That works fine with a battery. However, without a battery, the Pyra doesn't work and doesn't provide 1,8V there - so the Pyra thinks it's a low power device and limits it to 100mA, which is not enough to boot the system. There is a way to tell the Palmas (power management chip) using some software and a resistor to also accept more than 100mA. And yes, that worked, and the system boots up to U-Boot now. It doesn't start the kernel, which is something Nikolaus is looking into right now. That's some software tweak that needs to be implemented. This would've only been an issue if you're using the Pyra without a battery, but as some probably want to do that, we wanted to fix that. With a battery, it boots into the Linux kernel already, so don't worry about that. So we're moving ahead slowly, but we're pretty close to having a fully bootable system - though it's hard to tell how fast that will go, as some software changes require reading hundreds of pages in a datasheet and take days to solve whereas others are found within minutes. Still, it could very well be we've got a fully booting system when I leave for Greece - which would be a milestone, together with the case :) I'll keep you updated, as usual :) Want to leave feedback? Visit: https://pyra-handheld.com/boards/threads/the-flight-has-been-booked.76833/

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

Pyra-news January 2016